One area in computer-assisted self-interviewing (CASI) where methodological and empirical research have not caught up with technological advancements is real-time validation of respondent input. Some of the literature on computer-assisted interviewing and postsurvey editing is pertinent yet far insufficient—the largest component that could not be deduced being the respondent interaction. This leads to the employment or avoidance of validations by current practitioners based on untested assumptions. This article presents a model that breaks down the different elements in survey validation by type of computation, implementation, and interaction with the respondent. The purpose is twofold: to describe what possibilities exist in validations for web-based surveys and to present a framework in which systematic research could be conducted to evaluate the impact of validations on survey costs and errors. Relevant findings from prior literature are discussed in this context.
